At school, we might learn that glaciers erode by just two processes – abrasion, the grinding down of bedrock surfaces by debris carried on the glacier sole and plucking, the extraction of blocks from sockets on cliffs. And even how to pronounce roche moutonnée. But not that, for two centuries, geoscientists have overlooked another highly effective process of glacial erosion.

Ripped rock surface on granite gneiss, Gryttjen, Ljusdal, Gävleborg County, Sweden. (Photo: Adrian H
Ripped rock surface on granite gneiss, Gryttjen, Ljusdal, Gävleborg County, Sweden. (Photo: Adrian Hall).

There’s a lot of rubble in Sweden

Maps from the Swedish Geological Survey (SGU) show that dense concentrations of large, angular boulders litter vast areas in eastern Sweden. These boulder spreads extend for over 1000 km from Skåne to Lapland. The tracts of ankle-snapping rubble are hostile for farming and livestock, with few settlements and roads. Blocks are piled on rocks. The boulder spreads have been left mainly to forestry and so lie hidden in air photographs. The boulder faces scatter radiation so that even the latest high resolution airborne laser imagery fails to show the ground surface well. 

Adrian Hall, from the Department of Physical Geography at Stockholm University, takes up this ripping yarn: “To understand this stuff, you have to be out in the field, on the ground. There you find that in amongst all the boulders are small areas where the hard, ancient gneiss bedrock is more or less intact. But all the fractures have been opened up. On small rock hills, the whole rock surface has been ripped apart and started to move. At Bodagrottorna, Iggesund, we wriggled into short, dark and mazy caves between huge tilted monoliths. Some fracture caves reach down to 10 m below the surface.”

The disrupted surface of the large roche moutonnée at Bodagrottorna, Iggesund. The hill is riddled w
The disrupted surface of the large roche moutonnée at Bodagrottorna, Iggesund. The hill is riddled with fracture caves (Photo: Maarten Krabbendam).

The rock has been ripped apart

This impressive destruction has been explained before as the work of mega-earthquakes, equivalent in force to the 2010 Haiti quake, or of mysterious methane bursts. But mapping by structural and glacial geologist, Maarten Krabbendam, of the British Geological Survey, shows that the rock displacement is basically in one direction and that direction always corresponds to the last flow direction of the Fennoscandian ice sheet. The boulder spreads also have been moved down-ice from residual outcrops. So these features have a glacial origin.

A small roche moutonnée ripped apart by ice moving from right to left. Grundstugan, Lake Vallen, Upp
A small roche moutonnée ripped apart by ice moving from right to left. Grundstugan, Lake Vallen, Uppland (Photo: Maarten Krabbendam).

The late Robert Lagerbäck of SGU pointed the way as to how the bedrock might have been ripped apart by glaciers. Robert linked the broken bedrock to remarkable pictures from excavations at the nuclear power station site at Forsmark in the 1970s. There horizontal fractures in the gneiss bedrock had been prized open and filled with silt and sand. This was evidence of the development of massive water pressure below the last Fennoscandian ice sheet, pressure sufficient to hydraulically jack up not only the bedrock surface but also hundreds of metres of overlying ice.  

“So the search was on other signs of jacking and bedrock disruption”, reports Adrian. “And, as so often when you know what you are looking for, you find it. We had great fun checking rock quarries in Kalmar, Uppland and up towards Hudiksvall. There we found many rock faces with wide open fractures. And in their surroundings were extensive boulder spreads. Sometimes heaped into moraines.” 

Three steps to ripping

The whole assemblage points to a simple process sequence: jack, rip and transport. First, hydrostatic overpressure develops below the ice, where the entire weight of the ice sheet is applied to water confined between the ice and its bed or within fractures. This forces apart existing fractures, possibly extending them. A natural form of fracking, where water and sand are pumped between strata by drillers to open fractures and to maximise oil and gas yields.  With the gneiss in a temporary dilated and weakened state, wherever drag is applied by the ice moving across its bed, then the rock surface can be ripped apart.  This generates blocks which are then transported and deposited as boulder spreads.

How ripping works (Model diagram by Maarten Krabbendam, BGS)
How ripping works (Model diagram by Maarten Krabbendam, BGS).

Meltwater deluge

A key observation is that some distinctive boulders can be matched to rock outcrops nearby.  Typical transport distances are short, only 1 to 100 metres.  This proximity requires that the time available for glacial transport of boulders was very short. The implication is that the boulder spreads formed very close to the ice margin, at a time late in the retreat of the ice sheet. In Uppland this was about 10 000 years ago.  This makes sense because the melting ice generates huge volumes of subglacial meltwater.  The thinning ice also exerts less and less overburden pressure on the rock. This allows the rock surface to be mobilised.  This scenario mirrors current situations on the margins of the Greenland ice sheet, where man made climate change is forcing rapid melting.  The sudden drainage of lakes on top of the Greenland ice sheet down through moulins and crevasses transfers a deluge of meltwater directly to the ice sheet bed.  Recent borehole measurements of water pressures below the edge of the Greenland ice in SKB-sponsored research projects confirms that hydrostatic overpressure develops over periods of minutes to days during the melt season. For these brief periods, the rock surface and the ice above it are effectively floating.

So how important is this new process of glacial ripping?  The distribution of boulder spreads tells us that erosion by ripping in Sweden has operated over wide areas – but not everywhere.  The depth of boulder spreads is typically 1 to 4 m; this represents the depth of rock mobilised by ripping.  Recent modelling work in Uppland based on cosmogenic nuclides, by Jakob Heyman of Göteborg University, shows that the total depth of rock removed by the last ice sheet by abrasion and plucking was generally only about 0.6-1.6 m. So ripping operating right at the end of the last glaciations can remove as much rock than all the other processes of glacial erosion combined. And all that ripped rock and rubble is ready to be removed by next ice sheet.

A new window on the ice sheet bed 

The race is now on find out more. Like the rest of us, Adrian is keen to be out and about once we’re through Covid-19. “We have the bed of a former ice sheet to explore across Sweden and a new way to view it. The boulder spreads tell us where water was pressurised below the ice. We have a wonderful opportunity to better understand what happens hundreds of metres below ice sheets. And it’s not every day that you come across a new process of glacial erosion.” A ripping yarn indeed. 

The scientific article

Hall AM, Krabbendam M, Goodfellow BW, Hättestrand C, Heyman J, Palamakumbura RN, Stroeven A, Näslund J-O (In Press) Glacial ripping: geomorphological evidence from Sweden for a new process of glacial erosion Geografiska Annaler 2020
Read the article here.


Adrian Hall