The CIS hypothesis for jökulhlaups from inland ice sheets

The CIS hypothesis (Erlingsson 1994a, b, 2006) builds on the hydrodynamic formulas presented by Björnsson (1988, 2002) to explain the jökulhlaups of Vatnajökul. The principle is simple: To act as a dam, the ice needs to be higher on all shores around the subglacial lake. This “ice rim” creates a hydrostatic seal (Fig. 1a), provided that the ground below is impermeable.

Since a CIS has a very low slope, it closely follows the height contours. Unless the climate is so cold that there is accretion underneath, the seal will eventually break at some location around the perimeter. The jökulhlaup does not necessarily occur on the saddle point of the ground, since the ice rim elevation is ten times more important than the ground elevation for determining the efficiency of the hydrostatic seal. Nor does it necessarily happen on the water divide. A computer model (Erlingsson 1994b) predicted that it first happened beyond the water divide, and later before it.

After each jökulhlaup the ice shelf will stagnate, including any ice stream driven by the same pressure head. A wide zone of stranded CIS will melt away as dead ice, behind an end moraine deposited during the advance. The glacial geology will consist of more glaciolacustrine and glacifluvial deposits than typical for a classic ice sheet, and the diamicton will consist of glacitectonized varved clay rather than till.

If the ablation is insufficient the ice shelf will grow in thickness. The captured lake and the jökulhlaups will gradually diminish in magnitude during this transformation, and may finally cease altogether.

Since a CIS may be significantly thinner than a traditional ice sheet (Fig. 1b) it may bind less water, and cause less glacial isostacy. Furthermore, rain may be trapped in the captured lake without freezing, and re-released to the ocean in a jökelhlaup without having to be melted, which affects the energy budget for a glacial cycle.

On the other hand, as evidenced by Lake Vostok a CIS may also be very think, and on the surface appear as a near-horizontal interruption of the normal plastic flow of grounded ice. As illustrated in Figure A1c, the difference in volume compared to a grounded ice may be as little as 25% or less, even though a large captured lake is present. In this scenario, the ice sheet first melts downwards when the air temperature in the inland increases so that the base reaches the pressure melting point. A significant proportion of the meltwater accumulates under the ice, eventually to be released in the jökulhlaup associated with mwp-1a. The main difference between the CIS in Fig. 1a and 1b on the one hand, and 1c on the other, is the distance of it from the ice margin. The ice in front of it has to obey the normal flow law of ice, which determines the thickness of the ice downstream of the CIS. The thickness at the downstream end of the CIS has to be greater than the ice after the CIS.

The North American morphology allows an enormous CIS to form (Erlingsson 2008). The primary accumulation zone of the ice sheet is in the mountains east of Hudson Bay. Once the ice has grown so much that the bay becomes cut off from the ocean, meltwater and rain will flood a large part of the central continental lowland. An ice shelf spreading over it would become captured (land-locked) on all sides. The predicted location of the ice limit of a CIS (slightly beyond the water divide), matches the observed position of the Laurentian ice limit. Most jökulhlaups will have been discharged towards the Mississippi during the pleniglacial, and thus the Gulf of Mexico.

A consequence of the CIS theory is that repeated jökulhlaups may be expected. The frequency depends on the rate of accumulation of water in, and the volume of, the subglacial lake (Erlingsson 1994b). Thermal processes at the threshold may also lead to a cyclic development (Alley et al. 2005). The period of Laurentian jökulhlaups can be roughly estimated to several centuries to a few millennia, and the volume 0.3 to 3·1014 m3, based on the size of the basin, and an assumed jökulhlaup floating level drop in the order of 100 to 200 m. It would most likely pull with it large amounts of ice. One or several smaller CIS may have covered the Great Lakes.

In summary, landlocked ice sheets may contain large pulsating captured lakes, giving them quite different dynamics than the surviving sea-limited ice sheets of the Antarctic and Greenland. Since they are predicted to be preceded by an advance of low-slope ice lobes, and to be superseded by stagnation, we may search for them in the ice margin record. The southern margin of the Laurentian ice sheet contains a large number of end moraines that could reflect such events.

Read also about how memories of these events may have been preserved as the Lindorm and Unktehi in mythology.

Figure 1. A captured ice shelf lobe. (a) The hydrostatic seal under the ice rim is created where the equipotential level dips below impermeable ground. The equipotential level is a 10 times vertically exaggerated mirror image of the ice surface. The floating level represents the pressure head of the water emerging in a jökulhlaup when and where the hydrostatic seal is broken. (b) Virtual ice age photo up along the Missouri River from ca 41°N, 95°W. Created in Terragen software using ETOPO-2 data, with an ice lobe in the form of a CIS with top level 800 m a.s.l. added (see arrow). Vertical exaggeration 20 times. (c) Hypothetical section across the Laurentian ice sheet from ice divide to the southern margin, comparing an ice dome without water below, with an ice sheet featuring a large CIS below. Since the lake is far from the ice margin, the ice thickness is great.

Figure 2. Oblique aereal photo taken in a generally northern direction towards the ice lobes in the Missouri (left) and Mississippi (right). The rivers are flowing to the bottom right. The left lobe ends in Yankton County; note the wide floodplan of the Missouri beyond that point, which is pointed out as a possible location for the jökulhlaup in Erlingsson (2008). Pipestone, Minnesota, is located on the left side of the right lobe in this image, approximately due right from the tip of the left lobe.

References

Alley, R.B., Dupont, T.K., Parizek, B.R., Anandakrishnan, S., Lawson, D.E., Larson, G.J. and Evenson, E.B., 2005: Outburst flooding and the initiation of ice-stream surges in response to climatic cooling: A hypothesis. Geomorphology 75:76-89.

Björnsson, H., 1988: Hydrology of ice caps in volcanic regions. Societas Scientarium Islandica, XLV.

—2002: Subglacial lakes and jökulhlaups in Iceland. Global and Planetary Change, 35:255–271.

Dyke, A.S., Moore, A. and Robertson, L., 2003: Deglaciation of North America. Geological Survey of Canada, Open File 1574.

Erlingsson, U., 1994a: The ‘Captured Ice Shelf’ hypothesis and its applicability to the Weichselian glaciation. Geografiska Annaler 76A (1–2): 1–12.

—1994b: A computer model along a flow-line of an Ice Dome—‘Captured Ice Shelf’. Geografiska Annaler 76A (1–2): 13–24.

—2006: Lake Vostok behaves like a ‘captured lake’ and may be near to creating an Antarctic jökulhlaup. Geografiska Annaler 88A (1): 1–7.

—2008: A jökulhlaup from a Laurentian captured ice shelf to the Gulf of Mexico could have caused the Bølling warming. Geografiska Annaler 90A (2): 125–140. [ manuscript | paper ]

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buch, C.E., Burr, G.S., Cutler, K.B., Darnon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, G., Manning, S., Ramsey, C.B., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J. and Weyhenmeyer, C.E., 2004: IntCal04 Terrestrial Radiocarbon Age Calibration, 0–26 cal kyr BP. Radiocarbon, 46: 1029–1058.